FEMS Microbiology Letters 189 (2000) 311^315
www.fems-microbiology.org
Retention of bacteria on a substratum surface with micro-patterned
hydrophobicity
R. Bos a , H.C. van der Mei a , J. Gold b , H.J. Busscher
a
a;
*
Department of Biomedical Engineering, University of Groningen, Bloemsingel 10, 9712 KZ Groningen, The Netherlands
b
Department of Applied Physics, Chalmers University of Technology, S-412 96 Gotheborg, Sweden
Received 11 March 1999; received in revised form 15 May 2000; accepted 30 May 2000
Abstract
Bacteria adhere to almost any surface, despite continuing arguments about the importance of physico-chemical properties of substratum
surfaces, such as hydrophobicity and charge in biofilm formation. Nevertheless, in vivo biofilm formation on teeth and also on voice
prostheses in laryngectomized patients is less on hydrophobic than on hydrophilic surfaces. With the aid of micro-patterned surfaces
consisting of 10-Wm wide hydrophobic lines separated by 20-Wm wide hydrophilic spacings, we demonstrate here, for the first time in one
and the same experiment, that bacteria do not have a strong preference for adhesion to hydrophobic or hydrophilic surfaces. Upon
challenging the adhering bacteria, after deposition in a parallel plate flow chamber, with a high detachment force, however, bacteria were
easily wiped-off hydrophobic lines, most notably when these lines were oriented parallel to the direction of flow. Adhering bacteria detached
slightly less from the hydrophilic spacings in between, but preferentially accumulated adhering on the hydrophilic regions close to the
interface between the hydrophilic spacings and hydrophobic lines. It is concluded that substratum hydrophobicity is a major determinant of
bacterial retention while it hardly influences bacterial adhesion. ß 2000 Federation of European Microbiological Societies. Published by
Elsevier Science B.V. All rights reserved.
Keywords : Bio¢lm; Bacterial adhesion; Micropatterned surface; Hydrophobicity ; Bacterial detachment
1. Introduction
Over the past two decades, bacterial adhesion to substratum surfaces with di¡erent hydrophobicities has been
studied by an enormous number of research groups [1^4].
Whereas it is sometimes concluded that bacterial adhesion
is less to hydrophobic substrata [5], opposite conclusions
are also drawn [6]. The diversity in the bacterial world is a
factor contributing to this confusion, particularly since
bacterial cell surfaces can be as hydrophilic as glass while
other strains can have cell surfaces as hydrophobic as wax
[7]. Recently, however, we have become convinced that
this confusion is largely due to inadequate experimental
systems to study bacterial adhesion [8]. In a typical `Materials and methods' section of a paper on bacterial adhesion, it can be read that `substrata with adhering bacteria
were slightly rinsed under running tap water' or `dipped to
remove loosely adhering bacteria prior to enumeration'.
* Corresponding author. Tel. : +31 (50) 363-3140;
Fax: +31 (50) 363-3159; E-mail: [email protected]
Such methodology has been (and still is in many research
groups) general practice for many years, but nevertheless
is inadequate as it raises a number of obvious questions:
(a) what is the magnitude of the rinsing forces applied, (b)
what is the de¢nition of `loosely adhering' bacteria, (c)
what percentage of the total adhering population is removed by rinsing?
To avoid these questions, we have started doing bacterial adhesion experiments in a parallel plate £ow chamber
under controlled hydrodynamic conditions [8]. With the
aid of phase-contrast microscopy, ultra-long working distance objectives and image analysis, bacterial adhesion
could be observed `live', as it occurred and enumeration
was instantaneous under the shear conditions of the actual
experiment. Based on £ow chamber studies involving several bacterial strains and substratum materials, it became
increasingly di¤cult to conclude whether bacterial adhesion was more or less extensive on hydrophobic or hydrophilic substrata and it was often impossible to assign any
signi¢cance to trends observed.
As always, `the proof of the pudding is in the eating'
and for us the pudding was the human oral cavity and the
0378-1097 / 00 / $20.00 ß 2000 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 2 9 8 - 6
FEMSLE 9522 26-7-00
312
R. Bos et al. / FEMS Microbiology Letters 189 (2000) 311^315
oropharynx, two natural environments in which all surfaces exposed are prone to bacterial adhesion and bio¢lm
formation. We compared the formation of a dental bio¢lm
on hydrophobic strips glued to the front incisors of human
volunteers with bio¢lm formation on the more hydrophilic, natural dentition. We found that after 9 days refraining from all oral hygiene the hydrophobic strips were
virtually clean, while the natural enamel surface was 50^
75% covered with a bio¢lm [9]. Patients after total laryngectomy due to a malignant tumor in the head^neck area
are ¢tted with so-called shunt-valves, placed between the
trachea and the oesophagus, to help them speak [11]. The
oesophageal side of these prostheses becomes colonized
within 3^4 months with a thick bio¢lm, impeding proper
valve functioning. Valves, made partly hydrophilic and
partly hydrophobic, harvested a signi¢cantly thicker and
more extensive bio¢lm on their hydrophilic parts than on
their hydrophobic parts after being used for 4 weeks [11].
Finally, bio¢lm formation under the dynamic conditions
of the human oral cavity or the oropharynx is less on
hydrophobic substratum surfaces.
An evaluation of the di¡erences between our £ow chamber adhesion studies and the dynamic conditions in the
above natural environments led us to conclude that the
existence of £uctuating shear forces in the oral cavity
and in the oropharynx varies from being almost absent
to excessively high during eating, speaking and drinking.
Clearly this must be reproduced in the £ow chamber in
order to allow extrapolation of bacterial adhesion data to
the natural environment. It could be argued on the basis
of our combined in vitro £ow chamber experiments and in
vivo studies in humans that bacterial adhesion to substratum surfaces is not so much in£uenced by substratum
hydrophobicity, but by the retention of adhering bacteria
under £uctuating shear forces. The latter may be critically
determined by substratum hydrophobicity. One way to
mimic the occasionally high shear forces existing in vivo
is to pass an air-bubble through the £ow chamber. The
forces accompanying the passage of a liquid^air interface
can be calculated from theoretical models and amount to
some 1037 N on average and the method has previously
been employed in the semi-conductor industry to remove
micron-sized dust particles from silicone wafers [12].
The aim of this work was to demonstrate the in£uence
of substratum hydrophobicity on bacterial adhesion and
retention under controlled £ow and after exposure to a
well de¢ned shear force in one and the same experiment.
This would avoid as much as possible biological and other
experimental variations.
2. Materials and methods
2.1. Bacterial growth conditions
Streptococcus sobrinus HG1025 was cultured in Todd
Hewitt broth at 37³C in ambient air. For each experiment,
strains were inoculated from blood agar in a batch culture.
This culture was used to inoculate a second culture which
was grown for 16 h. Bacteria were harvested by centrifugation (5 min at 10 000Ug), washed twice with demineralized water and ¢nally suspended in a 10-mM potassium
phosphate solution (pH 6.8) to a concentration of 3U108
ml31 for use in the parallel plate £ow chamber.
2.2. Micro-patterning
Micro-patterned surfaces on glass were prepared by a
lithographic technique previously described by Healy et al.
[13]. Brie£y, cleaned glass plates were spin-coated with a
layer of photoresist (S-1813, Shipley, Marlborough, MA,
USA) and subsequently exposed to light through a photomask. Next, the exposed resist was dissolved away in an
alkaline solution and rinsed with water, baked at 100³C
and stored. Finally, the plates were coated with dimethyldichlorosilane (DDS) after which the undeveloped photoresist was removed by sonication in acetone. The result
was a chemical pattern consisting of 10-Wm wide, hydrophobic lines of DDS with 20-Wm hydrophilic spacings in
between. Water contact angles on similarly prepared DDS
coatings on glass were about 90³, while glass had a water
contact angle of around 20³. In addition, ellipsometry and
X-ray photoelectron spectroscopy indicated that the thickness of such a DDS coating was 1^4 nm [13]. The micropatterned surfaces could be inserted into the parallel plate
£ow chamber with the pattern either parallel or transverse
to a £owing bacterial suspension [10]. The position of the
hydrophobic lines was identi¢ed microscopically during
the actual adhesion experiment as described below.
2.3. Parallel plate £ow chamber
Deposition was observed on the bottom plate of the
parallel plate £ow chamber with a CCD-MXR camera
(High Technology, Eindhoven, The Netherlands) mounted
on a phase-contrast microscope (Olympus BH-2) equipped
with a 40U ultra-long working distance objective (Olympus ULWD-CD Plan 40 PL). The camera was coupled to
an image analyzer (TEA, Difa, Breda, The Netherlands).
Live images were Laplace-¢ltered after subtraction of an
out of focus image. Thereafter, adhering bacteria were
discriminated from the background by single gray value
thresholding. In this set up, one image covers a surface
area of 0.016 mm2 . The internal dimensions of the parallel
plate chamber were 3.8U5.6U0.06 cm (widthUlengthUheight) and £uid £ow was 1.5 ml min31 .
All experiments were carried out in duplicate with separately cultured bacteria and newly prepared patterned
surfaces. In each experiment, images were taken of three
di¡erent locations on the patterned surface prior to challenging the adhering bacteria with the high detachment
force exerted by a passing liquid^air interface, and of three
FEMSLE 9522 26-7-00
R. Bos et al. / FEMS Microbiology Letters 189 (2000) 311^315
di¡erent locations after the exposure of the adhering bacteria to a passing liquid^air interface. The liquid^air interface was created by introducing an air-bubble in the £ow
chamber, that completely spanned the width of the chamber.
The position of the hydrophobic lines could be discerned from the de-wetting pattern after the passage of
the liquid^air interface.
3. Results and discussion
Fig. 1 presents images of S. sobrinus HG1025 adhering
from a £owing suspension to patterned surfaces. In the
images taken prior to the passage of an air-bubble through
the £ow chamber, the micro-pattern can hardly be discerned, demonstrating that these streptococci do not
have a strong preference for adhesion to either hydrophobic or hydrophilic substrata. Retention of the adhering
streptococci, as observed in the images after their exposure
to a liquid^air interface, clearly reveals the micro-pattern
re£ected in the density of adhering bacteria able to withstand this detachment force with slightly fewer bacteria
adhering to the hydrophobic lines than to the hydrophilic
spacing and with a preferential retention of bacteria on the
313
hydrophilic spacings close to the interface between the
hydrophilic and hydrophobic regions. Interface regions
parallel to the £ow more distinctly reveal the micro-pattern from the density of adhering bacteria than lines transverse to the £ow.
Quantitative enumeration of the number of adhering
streptococci prior to and after exposure high shear-o¡
forces in Fig. 2 con¢rms the conclusion drawn from the
images presented in Fig. 1, and prior to the passage of a
liquid^air interface similar numbers of bacteria are found
adhering over the di¡erent substratum regions (on average
1.8U106 cm32 ). After the passage of a liquid^air interface
in a direction parallel to the patterning, up to 7U106 cm32
bacteria remain adhering on the hydrophilic spacings,
close to the interface region, while less than 0.5U106
cm32 are found on the hydrophobic lines. When the hydrophobic lines are transverse to the £ow and the direction
of liquid^air interface passage, the e¡ects of the micropatterned hydrophobicity on bacterial retention are less
pronounced.
The use of micro-patterned surfaces here for the ¢rst
time in one and the same experiment demonstrates that
bacterial adhesion is hardly in£uenced by substratum hydrophobicity, while substratum hydrophobicity is a major
determinant of bacterial retention. The observation that a
Fig. 1. Distribution of streptococci adhering on micro-patterned surfaces with hydrophobic lines on hydrophilic glass from a £owing suspension. Images
were taken after 4-h exposure of the surfaces to £ow; for the top panel (a and b), the hydrophobic lines were oriented parallel to the £ow, while for
the bottom panel (c and d), the lines were transverse to the £ow; the distribution shown in the left-hand panel was taken prior to exposure of the adhering bacteria to a high shear-o¡ force and the right-hand panel shows those bacteria that remained adhering after challenge by a liquid^air interface.
Arrows indicate 10-Wm wide hydrophobic lines.
FEMSLE 9522 26-7-00
314
R. Bos et al. / FEMS Microbiology Letters 189 (2000) 311^315
Fig. 2. Average quantitative distribution of adhering streptococci on micro-patterned surfaces. The arrangement of the graphs is similar as in Fig. 1,
with shaded areas indicating the hydrophobic lines. The standard deviation over six di¡erent locations, equally divided over two experiments, amounts
20%.
relatively large number of bacteria remain adhering close
to the interface between hydrophilic and hydrophobic regions after exposure to high shear forces is most interesting. This can possibly be explained by the detachment
forces arising from the passing liquid^air interface, which
more readily cause detachment from hydrophobic than
from hydrophilic regions due to di¡erent adhesive forces
and spontaneous de-wetting of the hydrophobic surface
[14]. Consequently, when the pattern is parallel to the
moving liquid^air interface, uninterrupted de-wetting occurs whereas in the alternative orientation de-wetting is
repeatedly interrupted resulting in two distinct bacterial
retention patterns. In addition, since bacteria are equipped
with a variety of di¡erent binding structures that can either be hydrophobic or hydrophilic, accumulation of bacteria near the interface between hydrophilic and hydrophobic regions might be a result of the most optimal
interaction between the heterogeneous bacterial cell and
the substratum surface. This speculation is supported by
observations that mammalian cells and also proteins preferentially accumulate on the most heterogeneous region of
wettability gradient surfaces [15].
The implications of this study for the development of
so-called `easy wipe-o¡' surfaces, as used in biomedical
engineering [16], the design of non-toxic paints to control
marine fouling on ship hulls [17] or in the prevention of
bio¢lm formation and biodeterioration of monumental
buildings are far reaching [18]. First of all, it emphasizes
a currently voiced opinion that bio¢lm models distinguishing only a dense, base bio¢lm and a more open-structured
surface bio¢lm [19] need to be extended with, by our definition, a linking ¢lm [20] (see also Fig. 3). The bacteria in
Fig. 3. Electron micrograph of a bio¢lm on a silicone rubber voice prosthesis with superimposed a bio¢lm model distinguishing a surface, base
and linking ¢lm, in which the linking ¢lm bacteria are the initially adhering bacteria in direct contact with the substratum. Bar indicates
50 Wm.
FEMSLE 9522 26-7-00
R. Bos et al. / FEMS Microbiology Letters 189 (2000) 311^315
the linking ¢lm are those that initially adhered to the
substratum surface and whose retention under periods of
high shear-o¡ forces is under direct control of the substratum surface hydrophobicity and the distribution of heterogeneous hydrophilic sites. These initially adhering bacteria link the entire bio¢lm growing on top of them to the
substratum and since their retention is governed by the
substratum surface, so is the retention of the entire bio¢lm. Based on this study, it can be concluded that development of easy wipe-o¡ surfaces for bio¢lm control in
biomedical and industrial applications with £uctuating
shear-o¡ forces operative is feasible, but research e¡orts
need to be re-directed from `bacterial adhesion' towards
`bacterial retention'.
[9]
[10]
[11]
[12]
[13]
References
[14]
[1] Van Loosdrecht, M.C.M., Lyklema, J., Norde, W. and Zehnder,
A.J.B. (1989) Bacterial adhesion: A physicochemical approach. Microbiol. Ecol. 17, 1^15.
[2] Costerton, J.W., Lewandowski, Z., Caldwell, D.E., Korber, D.R. and
Lappin-Scott, H.M. (1995) Microbial bio¢lms. Annu. Rev. Microbiol. 49, 711^745.
[3] Morra, M. and Cassinelli, C. (1997) Bacterial adhesion to polymer
surfaces : a critical review of surface thermodynamic approaches.
J. Biomater. Sci. Polym. Ed. 9, 55^74.
[4] Kitano, T., Yutani, Y., Shimazu, A., Yano, I., Ohashi, H. and Yamano, Y. (1996) The role of physicochemical properties of biomaterials and bacterial cell adhesion in vitro. Int. J. Artif. Organs 19, 353^
358.
[5] Bayer, R.E. (1982) Conditioning surfaces to suit the biomedical environment. J. Biomed. Eng. 140, 257^271.
[6] Olofsson, A.C., Zita, A. and Hermansson, M. (1988) Floc stability
and adhesion of green-£uorescent-protein-marked bacteria to £ocs in
activated sludge. Microbiology 144, 519^528.
[7] Van der Mei, H.C., Bos, R. and Busscher, H.J. (1998) A reference
guide to microbial cell surface hydrophobicity based on contact angles. Coll. Surf. B Biointerfaces 11, 213^221.
[8] Bos, R., Van der Mei, H.C. and Busscher, H.J. (1999) Physico-chem-
[15]
[16]
[17]
[18]
[19]
[20]
315
istry of initial microbial adhesive interactions ^ methods to study and
mechanism of. FEMS Microbiol. Rev. 23, 179^230.
Quirynen, M., Marechal, M., Busscher, H.J., Weerkamp, A.H.,
Arends, J. and Van Steenberghe, D. (1989) The in£uence of surface
free energy on planimetric plaque growth in man. J. Dent. Res. 68,
796^799.
Van den Hoogen, F.J., Oudes, M.J., Hombergen, G., Nijdam, H.F.
and Manni, J.J. (1996) The Groningen, Nijdam and Provox voice
prostheses: a prospective clinical comparison based on 845 replacements. Acta Otolaryngol. Stockh. 116, 119^124.
Everaert, E.P.J.M., Mahieu, H.F., Wong Chung, R.P., Verkerke,
G.J., Van der Mei, H.C. and Busscher, H.J. (1997) A new method
for in vivo evaluation of bio¢lms on surface-modi¢ed silicone rubber
voice prostheses. Eur. Arch. Otorhinolaryngol. 254, 261^263.
Leenaars, A.F.M. (1988) In: Particles on Surfaces 1: Detection, Adhesion and Removal (Mittal, K.L., Ed.), pp. 361^372. Plenum Press,
New York.
Healy, K.E., Thomas, C.H., Rezania, A., Kim, J.E., McKeown, P.J.,
Lom, B. and Hockberger, P.E. (1996) Kinetics of bone cell organization and mineralization on materials with patterned surface chemistry. Biomaterials 17, 195^208.
Gömez Suärez, C., Noordmans, J., Van der Mei, H.C. and Busscher,
H.J. (1999) Detachment of colloidal particles from collector surfaces
with di¡erent electrostatic charge and hydrophobicity by attachment
to air bubbles in a parallel plate £ow chamber. Phys. Chem. Chem.
Phys. 1, 4423^4427.
Ruardy, T.G., Schakenraad, J.M., Van der Mei, H.C. and Busscher,
H.J. (1997) Preparation and characterization of chemical gradient
surfaces and their application for the study of cellular interaction
phenomena. Surf. Sci. Rep. 29, 1^30.
Costerton, J.W. (1997) Cleaning techniques for medical devices: Bio¢lms. Biomed. Instrum. Technol. 31, 222^226.
Cooksey, K.E. and Wigglesworth-Cooksey, B. (1995) Adhesion of
bacteria and diatoms surfaces in the sea: a review. Aquat. Microbial.
Ecol. 9, 87^96.
Krumbein, W.E. (1996) Geophysiology and parahistology of the
interactions of organisms with the environment. Mar. Ecol. 17,
1^23.
Wimpenny, J.W.T. and Colasanti, R. (1997) A unifying hypothesis
for the structure of microbial bio¢lms based on cellular automaton
models. FEMS Microbiol. Ecol. 22, 1^16.
Busscher, H.J., Bos, R. and Van der Mei, H.C. (1995) Initial microbial adhesion is a determinant for the strength of bio¢lm adhesion.
FEMS Microbiol. Lett. 128, 229^234.
FEMSLE 9522 26-7-00